Zobrazeno 1 - 5
of 5
pro vyhledávání: '"Brinkmeyer, Lukas"'
Human motion prediction is a complex task as it involves forecasting variables over time on a graph of connected sensors. This is especially true in the case of few-shot learning, where we strive to forecast motion sequences for previously unseen act
Externí odkaz:
http://arxiv.org/abs/2212.11771
In fashion-based recommendation settings, incorporating the item image features is considered a crucial factor, and it has shown significant improvements to many traditional models, including but not limited to matrix factorization, auto-encoders, an
Externí odkaz:
http://arxiv.org/abs/2205.02923
Learning complex time series forecasting models usually requires a large amount of data, as each model is trained from scratch for each task/data set. Leveraging learning experience with similar datasets is a well-established technique for classifica
Externí odkaz:
http://arxiv.org/abs/2204.03456
The performance of gradient-based optimization strategies depends heavily on the initial weights of the parametric model. Recent works show that there exist weight initializations from which optimization procedures can find the task-specific paramete
Externí odkaz:
http://arxiv.org/abs/1910.12749
Autor:
Brinkmeyer, Lukas, Drumond, Rafael Rego, Scholz, Randolf, Grabocka, Josif, Schmidt-Thieme, Lars
Parametric models, and particularly neural networks, require weight initialization as a starting point for gradient-based optimization. Recent work shows that a specific initial parameter set can be learned from a population of supervised learning ta
Externí odkaz:
http://arxiv.org/abs/1909.13576